Fix some new pylint "E" warnings

After updating pylint, it started emitting additional "E"
warnings in some cases, fix them.

  unsubscriptable-object,
  unsupported-delete-operation

These were associated with the OVN AgentCache code. Instead
of using a subscript, create get/delete methods to do the
same thing.

  used-before-assignment

Re-factor some code so it's clear to pylint variables are
being assigned properly.

Trivialfix

Change-Id: I4a5ccb7f33465705e59b5274c41db3c371862b1e

.pylintrc

@@ -28,9 +28,6 @@ disable=
  no-method-argument,
  no-self-argument,
  not-an-iterable,
- unsubscriptable-object,
- unsupported-delete-operation,
- used-before-assignment,
 # "W" Warnings for stylistic problems or minor programming issues
  abstract-method,
  arguments-differ,

neutron/agent/linux/dhcp.py

@@ -391,10 +391,9 @@ class DhcpLocalProcess(DhcpBase, metaclass=abc.ABCMeta):
             with open(file_name, 'r') as f:
                 return converter(f.read()) if converter else f.read()
         except ValueError:
-            msg = "Unable to convert value in %s"
+            LOG.debug("Unable to convert value in %s", file_name)
         except IOError:
-            msg = "Unable to access %s"
-        LOG.debug(msg, file_name)
+            LOG.debug("Unable to access %s", file_name)
         return None
 
     @property

neutron/db/l3_db.py

@@ -1504,6 +1504,7 @@ class L3_NAT_dbonly_mixin(l3.RouterPluginBase,
             self._core_plugin, context.elevated(),
             {'port': port}, check_allow_post=False)
 
+        dns_data = None
         with plugin_utils.delete_port_on_error(
                 self._core_plugin, context.elevated(),
                 external_port['id']),\
@@ -1587,6 +1588,7 @@ class L3_NAT_dbonly_mixin(l3.RouterPluginBase,
             # raise the underlying exception
             raise e.errors[0].error
 
+        dns_data = None
         fip = floatingip['floatingip']
         with db_api.CONTEXT_WRITER.using(context):
             floatingip_obj = self._get_floatingip(context, id)

neutron/db/l3_hamode_db.py

@@ -503,6 +503,8 @@ class L3_HA_NAT_db_mixin(l3_dvr_db.L3_NAT_with_dvr_db_mixin,
         self._core_plugin.delete_network(admin_ctx, net.network_id)
 
     def safe_delete_ha_network(self, context, ha_network, tenant_id):
+        # initialize to silence pylint used-before-assignment warning
+        net_id = None
         try:
             # reference the attr inside the try block before we attempt
             # to delete the network and potentially invalidate the

neutron/plugins/ml2/drivers/ovn/agent/neutron_agent.py

@@ -195,7 +195,7 @@ class MetadataAgent(NeutronAgent):
         # If ovn-controller is down, then metadata agent is down even
         # if the metadata-agent binary is updating external_ids.
         try:
-            if not AgentCache()[self.chassis_private.name].alive:
+            if not AgentCache().get(self.chassis_private.name).alive:
                 return False
         except KeyError:
             return False
@@ -230,7 +230,7 @@ class OVNNeutronAgent(NeutronAgent):
         # If ovn-controller is down, then OVN Neutron Agent is down even
         # if the neutron-ovn-agent binary is updating external_ids.
         try:
-            if not AgentCache()[self.chassis_private.name].alive:
+            if not AgentCache().get(self.chassis_private.name).alive:
                 return False
         except KeyError:
             return False
@@ -257,7 +257,7 @@ class OVNNeutronAgent(NeutronAgent):
 
 
 @utils.SingletonDecorator
-class AgentCache:
+class AgentCache(object):
     def __init__(self, driver=None):
         # This is just to make pylint happy because it doesn't like calls to
         # AgentCache() with no arguments, despite init only being called the
@@ -273,7 +273,7 @@ class AgentCache:
         _agents = copy.copy(self.agents)
         return iter(_agents.values())
 
-    def __getitem__(self, key):
+    def get(self, key):
         return self.agents[key]
 
     def update(self, agent_type, row, clear_down=False):
@@ -286,7 +286,7 @@ class AgentCache:
             self.agents[agent.agent_id] = agent
         return agent
 
-    def __delitem__(self, agent_id):
+    def delete(self, agent_id):
         del self.agents[agent_id]
 
     def agents_by_chassis_private(self, chassis_private):

neutron/plugins/ml2/drivers/ovn/mech_driver/mech_driver.py

@@ -1359,7 +1359,7 @@ def get_agents(self, context, filters=None, fields=None, _driver=None):
 
 def get_agent(self, context, id, fields=None, _driver=None):
     try:
-        return n_agent.AgentCache()[id].as_dict()
+        return n_agent.AgentCache().get(id).as_dict()
     except KeyError:
         raise n_exc.agent.AgentNotFound(id=id)
 
@@ -1418,6 +1418,11 @@ def delete_agent(self, context, id, _driver=None):
         'SB_Global', '.', 'external_ids', delete_agent=str(id),
         if_exists=True).execute(check_error=True)
 
+    try:
+        n_agent.AgentCache().delete(id)
+    except KeyError:
+        LOG.debug('OVN agent %s has been deleted concurrently', id)
+
 
 def get_availability_zones(cls, context, _driver, filters=None, fields=None,
                            sorts=None, limit=None, marker=None,

neutron/plugins/ml2/drivers/ovn/mech_driver/ovsdb/ovsdb_monitor.py

@@ -327,7 +327,7 @@ class ChassisAgentDeleteEvent(ChassisAgentEvent):
             return False
 
     def run(self, event, row, old):
-        del n_agent.AgentCache()[row.external_ids['delete_agent']]
+        n_agent.AgentCache().delete([row.external_ids['delete_agent']])
 
 
 class ChassisAgentWriteEvent(ChassisAgentEvent):

neutron/profiling/profiled_decorator.py

@@ -42,41 +42,40 @@ def profile(f):
 
     @functools.wraps(f)
     def profile_wrapper(*args, **kwargs):
+        if not cfg.CONF.enable_code_profiling:
+            return f(*args, **kwargs)
 
+        profid = "%s.%s" % (f.__module__, f.__name__)
+        profiler = cProfile.Profile()
+        start_time = datetime.now()
         try:
-            if cfg.CONF.enable_code_profiling:
-                profid = "%s.%s" % (f.__module__, f.__name__)
-                profiler = cProfile.Profile()
-                profiler.enable()
-                start_time = datetime.now()
+            profiler.enable()
             return f(*args, **kwargs)
         finally:
-            if cfg.CONF.enable_code_profiling:
-                profiler.disable()
-                elapsed_time = datetime.now() - start_time
-                elapsed_time_ms = (elapsed_time.seconds * 1000.0 +
-                                   elapsed_time.microseconds / 1000.0)
-                stream = io.StringIO()
-                stats = pstats.Stats(profiler, stream=stream).sort_stats(
-                    SORT_BY)
-                stats.print_stats(cfg.CONF.code_profiling_calls_to_log)
-                stats.print_callers(cfg.CONF.code_profiling_calls_to_log)
-                stats.print_callees(cfg.CONF.code_profiling_calls_to_log)
-                profiler_info = utils.collect_profiler_info()
-                if not profiler_info:
-                    profiler_info = {'base_id': 'not available',
-                                     'parent_id': 'not available'}
-                LOG.debug('os-profiler parent trace-id %(parent_id)s '
-                          'trace-id %(trace_id)s %(elapsed_time)7d millisecs '
-                          'elapsed for %(method)s(*%(args)s, **%(kwargs)s):'
-                          '\n%(profiler_data)s',
-                          {'parent_id': profiler_info['parent_id'],
-                           'trace_id': profiler_info['base_id'],
-                           'elapsed_time': elapsed_time_ms,
-                           'method': profid,
-                           'args': args,
-                           'kwargs': kwargs,
-                           'profiler_data': stream.getvalue()})
-                stream.close()
+            profiler.disable()
+            elapsed_time = datetime.now() - start_time
+            elapsed_time_ms = (elapsed_time.seconds * 1000.0 +
+                               elapsed_time.microseconds / 1000.0)
+            stream = io.StringIO()
+            stats = pstats.Stats(profiler, stream=stream).sort_stats(SORT_BY)
+            stats.print_stats(cfg.CONF.code_profiling_calls_to_log)
+            stats.print_callers(cfg.CONF.code_profiling_calls_to_log)
+            stats.print_callees(cfg.CONF.code_profiling_calls_to_log)
+            profiler_info = utils.collect_profiler_info()
+            if not profiler_info:
+                profiler_info = {'base_id': 'not available',
+                                 'parent_id': 'not available'}
+            LOG.debug('os-profiler parent trace-id %(parent_id)s '
+                      'trace-id %(trace_id)s %(elapsed_time)7d millisecs '
+                      'elapsed for %(method)s(*%(args)s, **%(kwargs)s):'
+                      '\n%(profiler_data)s',
+                      {'parent_id': profiler_info['parent_id'],
+                       'trace_id': profiler_info['base_id'],
+                       'elapsed_time': elapsed_time_ms,
+                       'method': profid,
+                       'args': args,
+                       'kwargs': kwargs,
+                       'profiler_data': stream.getvalue()})
+            stream.close()
 
     return profile_wrapper

neutron/tests/functional/plugins/ml2/drivers/ovn/mech_driver/ovsdb/test_ovsdb_monitor.py

@@ -498,18 +498,18 @@ class TestAgentMonitor(base.TestOVNFunctionalBase):
 
     def test_agent_change_controller(self):
         self.assertEqual(neutron_agent.ControllerGatewayAgent,
-                type(neutron_agent.AgentCache()[self.chassis_name]))
+                type(neutron_agent.AgentCache().get(self.chassis_name)))
         self.sb_api.db_set('Chassis', self.chassis_name, ('other_config',
                 {'ovn-cms-options': ''})).execute(check_error=True)
         n_utils.wait_until_true(lambda:
-                neutron_agent.AgentCache()[self.chassis_name].
+                neutron_agent.AgentCache().get(self.chassis_name).
                 chassis.other_config['ovn-cms-options'] == '')
         self.assertEqual(neutron_agent.ControllerAgent,
-                type(neutron_agent.AgentCache()[self.chassis_name]))
+                type(neutron_agent.AgentCache().get(self.chassis_name)))
 
     def test_agent_updated_at_use_nb_cfg_timestamp(self):
         def check_agent_ts():
-            agent = neutron_agent.AgentCache()[self.chassis_name]
+            agent = neutron_agent.AgentCache().get(self.chassis_name)
             chassis_ts = self.sb_api.db_get(
                 'Chassis_Private', self.chassis_name,
                 'nb_cfg_timestamp').execute(check_error=True)
@@ -532,7 +532,7 @@ class TestAgentMonitor(base.TestOVNFunctionalBase):
         try:
             n_utils.wait_until_true(check_agent_ts, timeout=5)
         except n_utils.WaitTimeout:
-            agent = neutron_agent.AgentCache()[self.chassis_name]
+            agent = neutron_agent.AgentCache().get(self.chassis_name)
             chassis_ts = self.sb_api.db_get(
                 'Chassis_Private', self.chassis_name,
                 'nb_cfg_timestamp').execute(check_error=True)
@@ -541,14 +541,14 @@ class TestAgentMonitor(base.TestOVNFunctionalBase):
 
     def test_agent_restart(self):
         def check_agent_up():
-            agent = neutron_agent.AgentCache()[self.chassis_name]
+            agent = neutron_agent.AgentCache().get(self.chassis_name)
             return agent.alive
 
         def check_agent_down():
             return not check_agent_up()
 
         def check_nb_cfg_timestamp_is_not_null():
-            agent = neutron_agent.AgentCache()[self.chassis_name]
+            agent = neutron_agent.AgentCache().get(self.chassis_name)
             return agent.updated_at != 0
 
         if not self.sb_api.is_table_present('Chassis_Private'):